Re: virtcrypto_dataq_callback calls crypto_finalize_request() from irq context

From: Michael S. Tsirkin
Date: Thu Nov 02 2023 - 09:17:40 EST


On Thu, Nov 02, 2023 at 01:04:07PM +0000, Gonglei (Arei) wrote:
>
>
> > -----Original Message-----
> > From: Michael S. Tsirkin [mailto:mst@xxxxxxxxxx]
> > Sent: Wednesday, November 1, 2023 9:26 PM
> > To: Halil Pasic <pasic@xxxxxxxxxxxxx>
> > Cc: Gonglei (Arei) <arei.gonglei@xxxxxxxxxx>; Herbert Xu
> > <herbert@xxxxxxxxxxxxxxxxxxx>; Jason Wang <jasowang@xxxxxxxxxx>;
> > virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> > linux-crypto@xxxxxxxxxxxxxxx; Marc Hartmayer <mhartmay@xxxxxxxxxxxxx>
> > Subject: Re: virtcrypto_dataq_callback calls crypto_finalize_request() from irq
> > context
> >
> > On Sun, Sep 24, 2023 at 07:39:41PM +0200, Halil Pasic wrote:
> > > On Sun, 24 Sep 2023 11:56:25 +0000
> > > "Gonglei (Arei)" <arei.gonglei@xxxxxxxxxx> wrote:
> > >
> > > > Hi Halil,
> > > >
> > > > Commit 4058cf08945 introduced a check for detecting crypto
> > > > completion function called with enable BH, and indeed the
> > > > virtio-crypto driver didn't disable BH, which needs a patch to fix it.
> > > >
> > > > P.S.:
> > > > https://lore.kernel.org/lkml/20220221120833.2618733-5-clabbe@baylibr
> > > > e.com/T/
> > > >
> > > > Regards,
> > > > -Gonglei
> > >
> > > Thanks Gonglei!
> > >
> > > Thanks! I would be glad to test that fix on s390x. Are you about to
> > > send one?
> > >
> > > Regards,
> > > Halil
> >
> >
> > Gonglei did you intend to send a fix?
>
> Actually I sent a patch a month ago, pls see another thread.
>
>
> Regards,
> -Gonglei

And I think there was an issue with that patch that you
wanted to fix?
config changed callback got fixed but this still didn't.

--
MST